Truenas 12.2 SMB share keeps crashing

[thegreek1]

Hi,

I've been researching this error for a while now and I still haven't been able to find a solution. Each time I intiate a large data transfer from one location into the Samba share my server keeps powering off. There isn't much details in the /var/logs or dmesg output. Now I've moved from the gigabyte interface to a cheap line card and disabled hardware offloading. What else can be wrong as I can't find a solution for lock order reversal or the witness debugger.

root@truenas[~]# sysctl -a | grep hw.mem

root@truenas[~]#
root@truenas[~]# sysctl -a | grep hw.real
hw.realmem: 17179869184
root@truenas[~]#
root@truenas[~]# sysctl -a hw.model
hw.model: AMD FX(tm)-4100 Quad-Core Processor
root@truenas[~]# dmesg | grep ^CPU
CPU: AMD FX(tm)-4100 Quad-Core Processor (3624.40-MHz K8-class CPU)
CPU: AMD FX(tm)-4100 Quad-Core Processor (3624.40-MHz K8-class CPU)
root@truenas[~]#
root@truenas[~]#

Aug 16 14:00:51 truenas lock order reversal:
Aug 16 14:00:51 truenas 1st 0xfffff80076ecd2c8 dr->dt.di.dr_mtx (dr->dt.di.dr_mtx) @ /wrkdirs/usr/ports/sysutils/openzfs-kmod/work/zfs-f7ba541d64cbc60b21507bd7781331bea1abb12e/module/zfs/dbuf.c:3944
Aug 16 14:00:51 truenas 2nd 0xfffff8011322f820 dn->dn_struct_rwlock (dn->dn_struct_rwlock) @ /wrkdirs/usr/ports/sysutils/openzfs-kmod/work/zfs-f7ba541d64cbc60b21507bd7781331bea1abb12e/module/zfs/dbuf.c:3820
Aug 16 14:00:51 truenas stack backtrace:
Aug 16 14:00:51 truenas #0 0xffffffff80b4b7b1 at witness_debugger+0x71
Aug 16 14:00:51 truenas #1 0xffffffff80aec7b7 at _sx_slock_int+0x67
Aug 16 14:00:51 truenas #2 0xffffffff82c5353a at dbuf_check_blkptr+0x1da
Aug 16 14:00:51 truenas #3 0xffffffff82c521a1 at dbuf_sync_leaf+0x2a1
Aug 16 14:00:51 truenas #4 0xffffffff82c51a3b at dbuf_sync_list+0x3b
Aug 16 14:00:51 truenas #5 0xffffffff82c51d85 at dbuf_sync_indirect+0x295
Aug 16 14:00:51 truenas #6 0xffffffff82c51a87 at dbuf_sync_list+0x87
Aug 16 14:00:51 truenas #7 0xffffffff82c85da4 at dnode_sync+0xeb4
Aug 16 14:00:51 truenas #8 0xffffffff82c6514f at sync_dnodes_task+0x5f
Aug 16 14:00:51 truenas #9 0xffffffff82be769f at taskq_run+0x1f
Aug 16 14:00:51 truenas #10 0xffffffff80b3e2d8 at taskqueue_run_locked+0x168
Aug 16 14:00:51 truenas #11 0xffffffff80b3f254 at taskqueue_thread_loop+0x94
Aug 16 14:00:51 truenas #12 0xffffffff80aa2130 at fork_exit+0x80
Aug 16 14:00:51 truenas #13 0xffffffff80fd188e at fork_trampoline+0xe

 

[NugentS]

You don't seem to have any hardware other than 2 CPU's, a gigabyte inteface and a cheap line card.
I am going to take a guess at the following:

1. No memory
2. No disks
3. No HBA (or using a RAID card)
4. Some wierdo with AMD CPU's - there do seem to be a few of these and need specific BIOS settings

Followed by an actually sensible suggestion of "gigabyte interface" and "cheap line card". You should be using Chelsio or Intel and no crappy interface cards with potentially dubious drivers that tend to fail when doing this like large transfers or large quantities of simultaneous transfers at the same time - ie when pushing it

 

[thegreek1]

I think I found the answer I was looking for but didn't want to admit it, the issue is that both NIC cards are Realtek. Time to canabolize my other VM to use one of its spare NIC or the QLogic netextreme II card.